State of Art (SOTA) Report on Dense Ceramic Membranes for Oxygen Separation from Air Authors: Prof. Joe da Costa, The University of Queensland Dr Simon Smart, The University of Queensland Dr Julius Motuzas, The University of Queensland Prof. Shaomin Liu, Curtin University of Technology Prof. Dongke Zhang, University of Western Australia
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
State of Art (SOTA) Report on
Dense Ceramic Membranes for
Oxygen Separation from Air
Authors:
Prof. Joe da Costa, The University of Queensland
Dr Simon Smart, The University of Queensland
Dr Julius Motuzas, The University of Queensland
Prof. Shaomin Liu, Curtin University of Technology
Prof. Dongke Zhang, University of Western Australia
i
Executive Summary
Inexpensive, large scale oxygen production is crucial to the development of the next generation of
integrated carbon capture and sequestration power plants based on coal gasification or oxy-fuel coal
combustion technologies. The current state of the art is cryogenic distillation with no other mature
technology, such as pressure swing adsorption or polymeric membranes, able to cost effectively deliver the
tonnage quantities of O2 at the required purities of >95% (vol.). However, cryogenic distillation is an energy
and capital intensive mature technology with very few prospects for large performance improvements or
opportunities for cost reductions. Of the embryonic technologies, ion transport membranes (ITMs)
represent the most promising alternative. This research is primarily at the laboratory scale at present
where the state of the art are dense ceramic membranes made from doped perovskite materials such as
barium strontium cobalt iron mixed oxides (BSCF) which exhibit fluxes on the order of 10 – 15 ml min-1 cm-2
under optimal conditions. However, the more stable lanthanum strontium cobalt iron mixed oxides (LSCF)
is preferred owing to its superior chemical and mechanical properties, though delivering lower oxygen
fluxes. Fluorites are now being considered, particularly that these materials can operate under exposure to
water and CO2.
Commercial deployment has yet to be realised but small pilot scale facilities have been commissioned
using perovskites. The largest of these was built by Air Products and Chemicals in the USA with a
production capacity of 5 tonnes per day. This facility has been operational for in excess of 500 days, and its
success has undoubtedly led to their decision to construct a new 100 TPD facility due to be commissioned
in 2013. This is possibly the first indication that this technology can be taking over the hump in the
demonstration cycle (see Figure 1), where capital and operational costs could start reducing. The
membrane module design used by Air Products, a wafer-like module incorporating a central collection tube
arguably represents the state of the art by virtue of the success of their pilot facilities. Other researchers,
however, see future modules adopting tubular geometries to further reduce module size and membrane
sealing requirements. Regardless, the rate limiting steps for further commercialisation are the fabrication
of mechanically robust and chemically stable membranes (especially against water and CO2 exposure) and
the development of innovative sealing systems able to withstand the high operating temperatures required
as in both cases.
ii
Table of Contents 1 Introduction ............................................................................................................................................... 1
2.1 Research Drivers ................................................................................................................................ 1
Figure 1 – Maturity versus total investment for oxygen separation technologies.
Figure 2 - Dense ceramic oxygen conducting membranes (a) solid electrolyte for O2- conduction and (b)
mixed ionic electronic conductor.
Figure 3 – Ideal crystal structure of a perovskite compound.
Figure 4 – Schematic of O2 permeation through a perovskite membrane.
Figure 5 – Three-point and Four-point membrane modules. The box represents a membrane module and
the diagonal line in the box represents a membrane.
Figure 6 – Planar wafer ITM ceramic membrane stacks from Air Products & Chemicals (USA).
List of Tables
Table 1 - Typical Oxygen Requirements for ASU integration into CCS ready power plants
Table 2 - Summary of ASU technologies
Acknowledgements
The authors wish to acknowledge financial assistance provided through Australian National Low Emissions Coal Research and
Development (ANLEC R&D). ANLEC R&D is supported by Australian Coal Association Low Emissions Technology Limited and the
Australian Government through the Clean Energy Initiative. The authors would also like to acknowledge the insightful comments
and helpful suggestions from the two reviewers Prof. Terry Wall and Dr Paul Feron.
iv
List of Acronyms and Abbreviations
ANLEC Australian National Low Emissions Coal
ASU Air Separation Unit
BSCF Ba0.5Sr0.5Co0.8Fe 0.2O3-δ
CCS Carbon Capture and Sequestration (or Storage)
IGCC Integrated gasification combined cycle
ITM Ionic transport membranes
LSCF La0.5Sr0.5Co0.8Fe 0.2O3-δ
MIEC Mixed ionic electronic conducting
SOTA State of the art
TRL Technology Readiness Level
1
1 Introduction This report addresses the first deliverable for project number 3-0510-0034 Australian National Low
Emissions Coal Research and Development Ltd (ANLEC) title “Membranes for tonnage oxygen separation
suited to supply oxy-fuel and coal gasification applications”. This report was prepared by The University
of Queensland and project partners Curtin University and The University of Western Australia.
The objective of this report is on the state of art analysis (SOTA) of dense ceramic membranes for oxygen
separation from air. Hence, this report considers the three stages of technology development namely: (i)
embryonic, (ii) development, demonstration and deployment, and (iii) mature technologies. In this work,
we have assessed the open literature via university libraries, data base, scientific journals and available web
pages. Whilst there is a large number of information available, we have endeavoured to include only the
most relevant works that could be constituted as state of art in this report.
2 Background
2.1 Research Drivers Reducing the capital and operating costs of oxygen production is the major driver for the research and
development of ionic transport membranes (ITMs). This is particularly important for producing the tonnage
quantities of oxygen required for the next generation of integrated carbon capture and sequestration (CCS)
power plants, based on gasification and/or oxyfuel coal combustion processes. Currently large scale oxygen
production is carried out by cryogenic distillation, a mature technology where significant technological
innovations are no longer expected. Cryogenic distillation is capital and energy intensive, operating at very
low temperatures (-200oC) and at elevated pressures. Coupling a cryogenic air separation unit at the front
end of an oxyfuel coal power plant is likely to reduce power generation efficiencies by 25%, for example
taking a total power plant efficiency of around 40% down to 30%, which will, in turn, have profound
implications for the cost of electricity. Based on reputable work from the US Department of Energy, ITMs
offer the potential to separate oxygen from air and to reduce oxygen production costs by 35% or more1
with lower efficiency penalties.
2.2 Competitive Technologies There are multiple technology options available for oxygen production. The choice of technology used is
highly dependent on the type and scale of the application, final purity requirements and the cost of energy.
These technologies can be categorized according to their technology readiness level (TRL), grouped here
into (i) mature, (ii) DD&D (development, demonstration & deployment) and (iii) embryonic categories for
1 G.J. Stiegel, R.C. Maxwell, Gasification technologies: the path to clean, affordable energy in the 21st century, Fuel Processing Technology, 71
(2001) 79-97.
2
simplicity. Mature technologies offer a very small scope for further improvements as they have already
undergone major technical developments for at least the last 20 years. However, they offer the best
performance and risk minimisation profiles, and are therefore the most attractive for deployment in the
first generation of CCS plants. Embryonic technologies are those that are still at the laboratory research
stage. The gap between embryonic and maturity is herein referred to as DD&D, which represents those
technologies that have progressed beyond the laboratory proof-of-concept stage. DD&D technologies are
focussed on accelerating towards demonstration and deployment, but still must undergo further scientific
and engineering development. Hence, embryonic and DD&D technologies represent the largest scope for
improvement and efficiency gains but also the largest risk of failure.
Table 1 outlines the oxygen requirements for an ASU integrated with a carbon capture ready oxy-fuel or
IGCC power plant. It is important to realise that the oxygen purity requirements are dependent on the
combustion technology, CO2 capture technology and CO2 end use. A DOE NTEL study2 into pulverised coal
oxy-combustion technology found that if a high purity CO2 stream (> 95%) is required then it is more cost
effective to use a high purity O2 stream (99 vol%) rather than build additional capacity to remove the
contaminants from a lower quality O2 stream (95 vol%) during the CO2 clean up and compression trains.
However, for most cases the difference in cost and performance of using a higher purity or lower purity O2
stream was marginal.
Table 1 - Typical Oxygen Requirements for ASU integration into CCS ready power plants
O2 characteristic Requirement
Oxygen Flow rate 150 – 500 tO2 h-1
Oxygen Purity >95 vol% Allowed Impurities Ar < 5%
Figure 1 shows our opinion of the current stage of development for the most prominent and promising air
separation technologies. Table 2 provides more detailed information relating each technology including the
typical capital costs and energy requirements of the major air separation technologies discussed in this
section. Where possible we have used data where the chosen ASU technology has been integrated into
either an oxyfuel coal combustion or IGCC power plant. Currently only cryogenic air separation and ITMs
have undergone such a techno-economic evaluation and the data for polymeric membranes and PSA
systems are indicative of standalone systems only, making a true comparison difficult.
2 DOE/NETL, Pulverised coal oxycombustion power plants, volume 1: Bituminous coal to electricity. 2008.
3
Figure 1 – Maturity versus total investment for oxygen separation technologies.
Table 2 – Details of ASU technologies
Technology Cryogenic Air Separation
Pressure Swing Adsorption
Polymeric Membranes
Ionic Transport Membranes
Current Status Mature Mature Mature Demonstration
O2 Purity Limit (% vol) (Remaining impurities)
3
99+ (Ar)
95 (Ar)
40 (N2, CO2, H2O)
100
O2 Flowrate in largest commercial installations (tO2 d
-1)
>30004 <350
6 <20 for oxygen
enriched combustion applications
5 (soon to grow to 100 with installation of new plant)
Suitability for CCS High Moderate Alone – Low Hybrid - Moderate
5
High
Timeframe to CCS commercialization
Immediate Short NA Medium / Long
Installed Capital Cost in Oxyfuel or IGCC plant ($US2008 kWe
-1)
310 – 500*2,6
150 - 200 For standalone PSA system producing
<150 tO2 d-17
95 – 160 For a 30% O2 stream
8
260 - 2952,8
Energy Consumption in Oxyfuel or IGCC plant (kWh tO2
-1)
245 – 670*2,8
~4509-700
9 260 for 40% O2
stream10
190^ – 240
& for
hybrid membrane / cryogenic system
7
100 - 655#2,8
Energy Penalty for Integration with CCS
$
~25%2 NA NA ~25%
2
*the wide range in these values reflects the differing O2 flow rates required and energy integration opportunities available for oxy-fuel and IGCC applications #the largest value includes heating value of NG for generating required operating temperatures for ITM, whereas the smallest value represents only the electrical energy input $defined as 1-ηCCS/ηRef ^for plants producing > 1000 tO2 d
-1 &for plants producing < 100 tO2 d
-1
3 Smith, A.R. and Klosek, J., A review of air separation technologies and their integration with energy conversion processes. Fuel Processing Technology, 2001. 70(2): p. 115-134. 4 Castle, W.F., Air separation and liquefaction: recent developments and prospects for the beginning of the new millennium. International Journal of Refrigeration, 2002. 25(1): p. 158-172. 5 Burdyny, T. and Struchtrup, H., Hybrid membrane/cryogenic separation of oxygen from air for use in the oxy-fuel process. Energy, 2010. 35(5): p. 1884-1897. 6 DOE/NETL, Cost and Performance Baseline for Fossil Energy Plants Volume 3a: Low Rank Coal to Electricity: IGCC Cases. 2011. 7 Matson, S.L., et al., Membrane oxygen enrichment : II. Economic assessment. Journal of Membrane Science, 1986. 29(1): p. 79-96. 8 Bhide, B.D. and Stern, S.A., A new evaluation of membrane processes for the oxygen-enrichment of air. II. Effects of economic parameters and membrane properties. Journal of Membrane Science, 1991. 62(1): p. 37-58. 9 Meriläinen, A., Seppälä, A., and Kauranen, P., Minimizing specific energy consumption of oxygen enrichment in polymeric hollow fiber membrane modules. Applied Energy, 2012. 94(0): p. 285-294.
Research Development Demonstration Deployment Mature
4 Perovskites 5 Fluorites 6 Chemical looping materials
7 ITM modules 8 Air Products & Chemical ITM plant
Cu
rren
tFu
nd
ing
4
Mature technologies that overcame the “cost hump” for oxygen separation from air include cryogenic
distillation and pressure swing adsorption (PSA). For producing tonnage quantities of high purity (> 98 vol%)
oxygen, cryogenic distillation is the technology of choice whilst PSA systems find deployment at small to
medium scale or where the highest purity O2 is less of a concern10. Additionally the capital costs of PSA
systems tend to scale linearly with size making them a poor choice for integration with large power plants.
In the development phase, polymeric membranes typically deliver oxygen at purities less than 40%, and are
therefore used primarily to enrich oxygen or nitrogen streams. In the research phase, chemical looping
combustion has been proposed, though the development of stable metal oxides is in the early stages of
laboratory research.
The technology that is attracting major interest from the research community and subsequent investment
from government and industry sources is ionic transport membranes also known as dense ceramic
membranes. This technology spans from the embryonic stage, where there is still a concerted research
effort into developing new materials; to the developmental stage of membrane modules which are being
investigated in Europe, USA, Australia and Asia; and even the demonstration stage where Air Products and
Chemicals Inc. (USA) has operated a pilot plant for over 2 years, delivering 5 tonnes of oxygen per day (TPD)
and has recently begun construction of a 100 TPD facility in Convent, Louisiana with operation due to begin
in 201311.
3 Dense Ceramic Membrane Technology
3.1 Ionic Transport Materials The primary materials of interest for the synthesis of dense ceramics are perovskites (ABO3), fluorites (AO2),
brownmillerites (A2B2O5), Ruddlesden-Popper series (An+1BnO3n+1), and Sr4Fe6-xCoxO13 compounds12. These
materials conduct oxygen ions, essentially enabling oxygen separation from air. Perovskites are the most
attractive of these materials as they conduct oxygen ions and electrons spontaneously (Fig. 2b) and are
often referred to as mixed ionic electronic conductors (MIEC). The advantage of a MIEC material is that it
dispenses with the need for external electrical circuits (Fig. 2a) as is the case for fluorites, which operate
like fuel cells. Thus the use of MIEC materials simplifies the engineering design and reduces the operational
energy requirements for oxygen separation from air.
10 J.G. Jee, M.B. Kim, C.H. Lee Pressure swing adsorption processes to purify oxygen using a carbon molecular sieve. Chemical Engineering Science, 60 (2005) 869-882. 11
J.M Repasky, L. L. Anderson, VE. E. Stein, P. A. Armstrong, E. P. Foster, ITM Oxygen technology: scale-up toward clean energy applications,
International Pittsburgh Coal Conference, Pittsburgh, Pa., U.S.A. October 15–18, 2012 12
S. Smart, J.C. Diniz da Costa, S. Baumann, W.A. Meulenberg, Oxygen transport membranes: dense ceramic membranes for power plant
applications, in: A. Basile, S.P. Nunes (Eds.) Advance membrane science and technology for sustainable energy and environmental applications, Woodhead Publishing Ltd, Cambridge, 2011, pp. 255-294.
5
Figure 2 - Dense ceramic oxygen conducting membranes (a) solid electrolyte for O2- conduction and (b)
mixed ionic electronic conductor
3.1.1 Perovskites
Perovskite compounds are crystalline ceramics with a cubic structure13 described by the general formula
ABO3 (shown schematically in Fig. 3). The A and B sites are cations occupied by alkali or rare earth elements
of the Lanthanide series and transition metals, respectively. The unit cell is a face-centred cubic crystal with
the larger A cations located at the corners, the smaller B cation located in the body-centred position and
the O2- anions located in the face-centred positions. At high temperatures (>700 °C) they conduct oxygen
ions which diffuse through the crystal lattice via oxygen vacancy sites or defects14. The pioneering work of
Teraoka15 and co-workers demonstrated that vacancy defect concentration (), and consequently ionic
diffusion, was enhanced by doping the cation sites (A or B) with other cations (A’ or B’) of different sizes
and/or valences, resulting in the general formula AxA’1-xByB’1-yO3-δ.
Ba0.5Sr0.5Co0.8Fe 0.2O3-δ (BSCF) has been the most studied perovskite as it has consistently demonstrated the
highest oxygen fluxes of all the MIEC materials at ~3 ml min-1 cm-2 for discs and ~9 ml min-1 cm-2 for hollow
fibres at around 900oC with a transmembrane O2 partial pressure
differential of ~20 kPa. As with all membrane technologies, the
thickness of the membrane significantly influences the overall
oxygen flux. Disc membranes tend to be thicker (~1mm) and
therefore deliver lower oxygen fluxes than the thinner hollow fibres
(~0.3 mm). However, BSCF is unstable at temperatures below ~825
°C leading to crystal phase change and membrane mechanical
failure. This is significant as lower operating temperatures are more
desirable to reduce additional heating requirements. Hence, the more stable La0.5Sr0.5Co0.8Fe 0.2O3-δ (LSCF) is
generally preferred, despite its lower oxygen fluxes. LCSF hollow fibres generally give oxygen fluxes of 1 ml
J. B. Goodenough, Ceramic technology - Oxide-ion conductors by design, Nature, 404 (2000) 821. 15
Y. Teraoka, H. M. Zhang, S. Furukawa, N. Yamazoe, Oxygen permeation through perovskite-type oxides, Chemistry Letters, 11 (1985) 1743.
Figure 3 – Ideal crystal structure
of a perovskite compound.
A
O
B
6
min-1 cm-2, although these membranes have been
reported to operate for over 1000 hours. Replacing
strontium in BSCF with zirconia to form BCFZ has resulted
in more stable membranes which have been operated for
up to 2200 hours16. Crucially for oxy-fuel coal combustion,
perovskites are inherently unstable when exposed to CO2
which reacts with perovskites components to form non-
conducting carbonates.
3.1.2 Fluorites
Fluorites are pure ionic conductors, and whilst they can also separate pure oxygen from air, their lack of
electronic conductivity necessitates an external circuit to enable oxygen production, as depicted in Figure
2a. Several fluorite oxides such as yttria-stabilized zirconia (YSZ), samarium-doped ceria (SDC) or
gadolinium-doped ceria (GDC) do possess sufficient chemical stability for use in ITM applications as these
are widely applied as electrolyte layers in solid oxide fuel cells (SOFC)11. Many of the SOFC operation
involve exposure of these pure ionic conductors to CO2 and H2O. However, oxygen fluxes of fluorites are
generally extremely low, on the order of less than 0.05 ml min-1 cm-2. The only exception is bismuth
samarium doped ceria (BiSDCSm)17 based fluorites with oxygen fluxes as high as 0.9 ml min-1 cm-2 at 800oC.
3.2 Oxygen transport through membranes Oxygen diffusion through dense ceramic membranes can occur only if there is a driving force which
physically manifests as an oxygen partial pressure difference between the feed stream and permeate
stream. In terms of membrane operation, oxygen transport involves five progressive steps shown
schematically in Figure 418:
Step 1 - feed side gas transport: the oxygen molecule is transported from the gas phase to the
membrane surface by gas to gas diffusion.
Step 2 - dissociation (surface reaction) on interface I (feed side): the oxygen molecule adsorbs to the
membrane surface and then disassociates due to catalytic activity of the ceramic material.
Step 3 - ionic transport (bulk diffusion): the oxygen ions diffuse through the ceramic crystal lattices,
driven by a partial pressure gradient of oxygen across the membrane. Electrons are transported in the
opposite direction to maintain electrical neutrality of the membranes.
Step 4 - association (surface reaction) on interface II (permeate side): the oxygen ions recombine into
oxygen molecules and desorb from the membrane surface.
16 T. Schiestel, M. Kilgus, S. Peter, K. J. Caspary, H. Wang , J. Caro, Hollow fibre perovskite membranes for oxygen separation. Journal of Membrane Science, 258 (2005) 1-4. 17
K. Zhang, Z. Shao, C. Li, S. Liu, Novel CO2-tolerant ion-transporting ceramic membranes with an external short circuit for oxygen separation at
intermediate temperatures, Energy & Environmental Science, 5 (2012) 5257-5264. 18
S. Smart, J.C. Diniz da Costa, S. Baumann, W.A. Meulenberg, Oxygen transport membranes: dense ceramic membranes for power plant
applications, in: A. Basile, S.P. Nunes (Eds.) Advance membrane science and technology for sustainable energy and environmental applications, Woodhead Publishing Ltd, Cambridge, 2011, pp. 255-294.
Figure 4 – Schematic of O2 permeation
through a perovskite membrane.
7
Step 5 - permeate side gas transport : the oxygen molecules are transported to the permeate stream by
gas to gas diffusion.
Perovskite materials are inherently catalytic and able to break down molecular oxygen into oxygen ions at
high temperature. As a result the O2 fluxes in steps 2 and 4 are controlled by the kinetics of the surface
dissociation/association reaction, whilst the flux in step 3 is controlled by bulk diffusion associated with the
membrane thickness. The oxygen flux through the entire membrane is therefore limited either by surface
kinetics or by bulk diffusion. Thus the membrane materials and fabrication geometry become intrinsically
important. ITM are similar to any other membrane technology in the sense that by reducing the thickness
of the membrane, gas permeation or O2 flux increases likewise. In other words thinner membranes are
desirable as they deliver high oxygen throughput per area of membrane. However, a major difference of
ITMs is that there is a critical length (Lc) where any further reduction in membrane thickness no longer
increases O2 flux. At this critical length, the transport resistances due to surface kinetics and bulk diffusion
are equal, and the transport of oxygen becomes limited by the kinetics of the surface exchange reaction.
The Lc parameter is material and temperature dependent in the case of BSCF membranes, and it is
approximately 0.7 - 1.1mm between 800 - 900oC19. Once the membrane thickness is less than Lc, the
membrane transport is controlled by the kinetics of the surface reaction. Researchers have deposited
palladium or platinum catalysts on the surface of thin MIEC hollow fibres (200 - 300 μm thick) to overcome
surface kinetics limitations. In a recent study that employed palladium catalysts on BSCF hollow fibres
researchers observed that oxygen fluxes could be increased by a factor of 10 at lower temperatures (700
°C) and recorded a maximum of 14.5 ml min-1 cm-2 at 950 °C20. In the case of fluorites, the oxygen flux is
controlled by the electronic conductivity as the materials are ionic conductors. Nevertheless, as the
membrane thickness is decreased beyond Lc, the controlling step will no longer be bulk diffusion but the
kinetics of the surface exchange reaction21.
4 Engineering development phase Two types of design are being considered in the initial development phase of ITM modules as schematically
shown in Figure 5. The primary difference between the operating modes is how the driving force, i.e.
oxygen partial pressure difference, is established between the feed and permeate streams. Both seek to
maximise this driving force, in different ways, as the higher the driving force, the larger the oxygen flux.
19
W.K. Hong, G.M. Choi, Oxygen permeation of BSCF membrane with varying thickness and surface coating. Journal of Membrane Science 346,
(2010) 353-360. 20
A. Leo, S. Smart, S. Liu, J. C. Diniz da Costa, High Performance Perovskite Hollow Fibres for Oxygen Separation, Journal of Membrane Science, 368
(2011) 64-68. 21
H.J.M. Bouwmeester, H. Druidhof, A.J. Burggraaf, P.J. Gellings, Oxygen semipermeability of erbia-stabilized bismuth oxide. Solid State Ionics, 53-
56 (1992) 460-468.
8
Figure 5 – Three-point and Four-point membrane modules. The box represents a membrane module and
the diagonal line in the box represents a membrane.
The first type is called a three-point membrane module. In this set up, air is cleaned and dehydrated before
reaching the membrane module. Hence, the feed side of the membrane will contain pure air, possibly with
very small concentrations of water and CO2 (at ppm levels or lower). In this set up, (i) the air feed stream
could be pressurised in excess of 5 bars and the permeate side be maintained at 1 bar, (ii) the air feed
stream could be kept at 1 bar and the permeate stream at vacuum pressure, or (iii) for maximum driving
force the air feed stream could be pressurised and the permeate stream kept under vacuum pressure.
The second type is called a four-point membrane module which is envisioned for oxyfuel coal power plants.
In this set up, a highly concentrated CO2 stream (recycled from the combustion gases) is used as a sweep
gas stream on the permeate side. The function of the CO2 sweep gas is to reduce the partial pressure of O2
in the permeate stream and thereby increase the transmembrane O2 partial pressure driving force. This
serves the dual purpose of reducing the need to pressurise the feed stream and acts as a diluent for the
combustion gas to better control flame temperature in the burner. Hence, there is less of a need to
pressurise the air feed stream to high pressures, although the membranes employed in this set up must, by
definition, be chemically stable to water and CO2.
4.1 Materials selection
4.1.1 Three-point Membrane Module
Currently LSCF is the state of art material for application in the three-point membrane module as it
represents the best compromise between good oxygen fluxes (around 1 ml min-1 cm-2) and long term
stability. Of course, membranes with higher fluxes would be preferred, particularly since higher fluxes will
reduce the requirements for membrane area, corresponding unit size, and ultimately capital costs of the
K. Zhang, Z. Shao, C. Li, S. Liu, Novel CO2-tolerant ion-transporting ceramic membranes with an external short circuit for oxygen separation at
intermediate temperatures, Energy & Environmental Science, 5 (2012) 5257-5264. 24 Liu, S. and Gavalas, G.R., Preparation of oxygen ion conducting ceramic hollow-fiber membranes. Ind. Eng. Chem. Res., 2005. 44: p. 7633-7637.
10
especially under the vibration and frictional stresses experienced in an industrial environment. Therefore, it
is expected that the optimal compromise will involve membranes capillaries (1< Ø <10 mm) or tubes (Ø>10
mm) and replace hollow fibres as state of the art. Indeed, recent work at both The Fraunhofer Institute of
Ceramic Technologies and Systems (IKTS) and RWTH Aachen have seen both groups independently produce
larger scale membrane modules (up to 1m2 in membrane area) utilising tubular geomteries25. Flat
geometries are typically problematic to handle due to the inherent brittle nature and increased sealing
requirements and so have typically only been constructed in laboratory settings at very small dimensions.
However, it must be noted that the most successful demonstration of ITM technology has been the Air
Products design which utilises a variation of flat sheet technology in its 5 TPD plant.
To increase the mechanical robustness of dense ceramic membranes, there have been several attempts to
coat a dense perovskite layer on more mechanically robust porous tubular supports. The theory being that
the porous support would allow oxygen to access the perovksite layer which also provide the mechanical
stability for the thin perovskite layer. However, the thermal expansion coefficient of perovskites is highly
material specific and frequently non-linear and the membranes fabricated generally suffer from high air
leakage. This problem has been tackled by several groups around the world, but remains to be solved. In
the case of some fluorites26, the thermal coefficient expansion tends to be linear. Hence, fluorites could be
more attractive for coating on porous structures of mechanically robust materials. Currently, the state of
the art regarding mechanical strength is produce membranes from a single material only, with thicker
membrane dimensions, which has the natural trade-off of reducing oxygen production.
4.3 Membrane Module Engineering Challenges One of the great advantages of membrane systems is their traditional modularity which reduces process
complexity and enables relatively simple expansion (or reduction) of production capacity. It is expected
therefore, that an air separation unit utilising ITM technology would be modular in design and will likely
contain a large number of membrane modules. Until membrane fluxes of the more stable perovskites are
improved this will likely represent a significant capital investment, but on the other hand it should enable
operational flexibility which may decrease operational costs. The major engineering problem is sealing
membranes in such a way as to allow continuous operation at high temperature (up to 1000 °C) with
leakage rates of <10%. The primary engineering focus to reduce the leakage rates has been to reduce the
sealing area. In this case, hollow fibres, capillaries and tubes are preferred over flat geometries.
25
X. Dong, W. Jin, Mixed conducting ceramic membranes for high efficiency power generation with CO2 capture. Current Opinion in Chemical
Engineering , 1 (2012) 163-170. 26
S. Smart, J.C. Diniz da Costa, S. Baumann, W.A. Meulenberg, Oxygen transport membranes: dense ceramic membranes for power plant
applications, in: A. Basile, S.P. Nunes (Eds.) Advance membrane science and technology for sustainable energy and environmental applications, Woodhead Publishing Ltd, Cambridge, 2011, pp. 255-294.
11
The primary concern here is again related to matching the thermal expansion coefficient of the sealant with
the dense ceramic membranes and the module itself. To address this problem, either special sealing
compounds are required or the membrane/seal interface must be designed outside the high temperature
region (~1000°C) region of the membrane module. This allows for the membrane / seal interface to be kept
relatively cooler, at temperatures <600°C, where engineering sealing solutions become more technically
and economically feasible. This has the added benefit of reducing any potential chemical interactions at the
membrane / seal interface, favouring long term operation. At the moment there are no state of the art
seals for dense ceramic membranes commercially available and research groups typically employ a
combination of noble metals and heat reduction techniques. Proprietary seals, such as those used by Air
Products target the dense ceramic material; much is the same manner as the seals in high temperature
solid oxide fuel cells.
There is only one membrane module fully reported in the literature by Li’s group based on hollow fibre
geometry27. The membrane module consisted of 889 hollow fibres totalling 9914 cm2. One end of the
hollow fibre was sealed by the same ceramic membrane material as the hollow fibre, while the other end
was left open for the flow of oxygen. A low-temperature silicone sealant was used (maximum 350oC).
Hence, the sealed area was isolated from the membrane module high temperature region. The membrane
module operated at 1070 °C and delivered a maximum 3.1 l (STP) min−1 with the purity of 99.9%. This
designed reduced the effective membrane area as only the hollow fibres were effective in oxygen air
separation only at high temperatures. A similar approach has been shown in conferences by RWTH Aachen
University (Germany) where a fundamental mechanical design is under development of perovskite tubes
and seals. The German design includes a water cooling system around the sealing area of tubes.
5 Demonstration The state of the art in demonstration of dense ceramic membranes for oxygen separation from air has been
achieved by Air Products & Chemicals in the USA. This is the only demonstration plant at industrial scale
with proof of concept for long term operation, delivering 5 TPD of oxygen for over 515 days28. This major
achievement by Air Products & Chemicals was significantly funded by the USA Department of Energy, on
the order of $148 million since 1999 to scale up the technology29. The approach taken by Air Products
involved an innovative module design. They developed flat membranes in a wafer configuration as depicted
27
X. Tan, Z. Wang, B. Meng, X. Meng, K Li, Pilot-scale production of oxygen from air using perovskite hollow fibre membranes, Journal of
Membrane Science, 352 (2010) 89–196. 28
S. Smart, J.C. Diniz da Costa, S. Baumann, W.A. Meulenberg, Oxygen transport membranes: dense ceramic membranes for power plant
applications, in: A. Basile, S.P. Nunes (Eds.) Advance membrane science and technology for sustainable energy and environmental applications, Woodhead Publishing Ltd, Cambridge, 2011, pp. 255-294. 29
P. Armstrong, K. Fogash, Oxygen Production Technologies: Cryogenic and ITM, 2nd Int’l Oxy-Combustion Workshop, Windsor, CT, USA, Jan 25-26,